The SN25P comes with the mobo already installed - a SFF mobo. He'll need to post on Shuttle's website or - even better - call Shuttle and get some tech support. I wonder if that PSU is having issues with the load that is being placed on it....350W and that system still makes me cringe.
<font color=red>You're a boil on the arse of progress - don't make me squeeze you!</font color=red>
Guys, thanks for the replies. There are several of us with similar problems that have been comparing notes in another forum. We've all done extensive, recursive troubleshooting with up-to-date drivers, BIOS, etc. (and back-reving too) and have open cases with Shuttle tech support.
What sparked my question in this forum is the dual-core stress test article. Yes, my system will run benchmarks in dual-core mode till the cows come home. But busy out the USB and it's a different story.
So let me ask it this way: Do any of the "stress tests" that THG performs include stressing out the USB? I don't see any way to ask the THG guys directly, hence my inquiries here.
Frequent system hangs or crashes to black screen in dual-core mode (no BSOD, event logs or crashdumps). Only external indication is Link and Power LED's on wireless-G device turn off. Wireless keyboard/mouse USB receiver LED still shows activity when mouse is moved and SN25P power LED remains on. SN25P will respond to Reset button and Power button. Fans and disks remain on.
During periods of high USB activity in dual-core mode, for example, simultaneous wireless-G USB file transfer plus video capture in background, hang/crash is within seconds. Video capture alone can cause hangs within minutes. Wireless-G USB alone can cause hangs in minutes to hours, depending on activity. Wireless keyboard/mouse alone can cause hangs within hours.
In dual-core mode, two instances of Prime95 small torture test (with affinity for each core) will run indefinitely. CPU temp remains under 50C (21C at idle with Cool n Quiet enabled). Memtest86 will run indefinitely. SATA Test runs successfully with no errors reported. Aquamark and 3DMark run repeatedly with no errors. GPU temp remains under 72C.
NO stability problems if Windows is booted with /onecpu switch (all USB devices work together). NO stability problems with Athlon 64 3000+ installed. All USB devices work together connected to a laptop with no problems. Another workaround is to configure USB as version 1.1 only.
Windows XP with SP2 has been clean installed multiple times. There has never been a hiccup or problem during the installation process.
Shuttle SN25P Motherboard V1.2
BIOS version fn25s01o
Athlon 64 X2 4400+ (OPN# ADA4400DAA6CD)
2 GB DDR400 (Kingston KVR400X64C3AK2/2G)
Leadtek PX6600GT TDH (GeForce 6600GT)
(2) 74 GB WD Raptor SATA in RAID 0
Sony DRU-720 DVD+-RW (internal IDE)
HP DC4000 Video capture device w/Sony DRU-700 +-RW (USB)
Hawking HWU54D Hi-Gain Wireless-G (USB)
HP f1905 19" LCD Monitor
HP Wireless Keyboard/Mouse (USB)
WinXP Pro SP2
========== Driver Version Information ==========
Audio Driver 5.12.01.3645
Graphics Driver 77.72 (184.108.40.20672)
Ethernet Driver 4.75 (1.00.00.0475)
IDE Driver 5.18 (5.10.2600.0518)
Wireless NIC Driver 220.127.116.11
Nexus 7 are you using two 1gb memory modules, 4 512mb modules? In your system.
Have you tried reducing your memory to 1gb and try the tests?(Obviously depending on memory configuration you have) Possbile power supply problem with the increase load using two cpu's and USB just brings out the problem. Looks like you have effectively isolated condition but not anyone component. Any bios settings to control ram timings etc. have you tried?
<P ID="edit"><FONT SIZE=-1><EM>Edited by noko on 07/14/05 09:05 PM.</EM></FONT></P>
2 x 1GB DDR400. Yes, I've tried running with just one. Even tried with just one disk drive. Memory is timed at 3-3-3-8-1 per Kingston datasheet. All other timings are bone stock as well, no over clocking. Tried lowering multiplier to get core speeds of 2.0 and 1.8 GHz to no avail. Tried bumping memory, chipset and CPU voltages to upper limit of spec to no avail.
Three conditions must exist:
1) Dual-cores under WinXP SP2
2) USB2 (actually 1.1 & 2) set in BIOS
3) High-speed USB2 device connected
Remove any one of the conditions and the freezes do not occur.
BTW, I forgot to mention that, in addition to Memtest86 and Prime95 running indefinitely, Knoppix 3.9 is solid as a rock.
It really doesn't sound like a hardware problem, reason being all test you performed seems ok. If the USB had a hardware problem it would have shown up with the single core cpu. Soundls like to me a driver problem particulary with the USB port, maybe a hardware conflict is formed once you enable both cores. Well if you have a slot left, another option is to try out a USB2 card and see if the problems continues, I bet they won't.
As a side note it is nice seeing someone using 1gb modules and it looks to be ok at specs speeds. Keep us up to date.
<P ID="edit"><FONT SIZE=-1><EM>Edited by noko on 07/17/05 01:37 AM.</EM></FONT></P>
Yes, the problem is positively reproducible with both cores running and absolutely non-existent with only one core running. With no high-speed USB ("USB2") peripherals mounted, the thing is solid as a rock, even with a mild overclock (10% CPU and 5% GPU).
A Shuttle tech (linuxnutbox) has been cool enough to join us on his own time over at the Sudhian SFFTech forum to understand what we've been dealing with. He has informed us that our problem got visibility with their CTO and now there are a couple of BIOS engineers working on resolving the problem.
I was just posting here to see if anyone else running the nForce4 Ultra chipset (not necessarily in an XPC) was having similar problems with X2's and USB2. Sounds like maybe not, which on the bright side might actually mean that a patch to the SN25P's BIOS might actually fix it.
Believe it or not your post was good news, in that the X2 handles 1gb dimms at pc3200 speeds without to much problem. Now I would like to know if the X2 will handle 4gb with four dimms at pc3200 speeds which would make the X2 more appealing to me. More later, please let us know when this is resolved or not.